The Graduation Proficiency Test in the College of Liberal Arts,
University of Minnesota

The Graduation Proficiency Test (GPT), which is used to measure language proficiency at the level required for University students who earn undergraduate degrees in the College of Liberal Arts, has been in use since 1986. Information about the test and the language requirement is available from a number of sources; this document summarizes key policies and committee findings having to do with the education requirement as it pertains to students at the university, test design, and state-wide articulation efforts. It has been prepared as a working document by members of the Committee on Second Language Education (ComSLE) for use at meetings where the GPT and language requirement are to be discussed.

The Educational Requirement

The University of Minnesota Mission Statement emphasizes a commitment to connecting education on campus with "...the application of this knowledge to benefit the people of the state, the nation, and the world." (CLA Bulletin, 2) This commitment is further elaborated in connection with the research, teaching, and outreach missions of the University. The research and discovery conducted here are to "benefit students, scholars, and communities across the state, the nation, and the world"; teaching and learning are undertaken to prepare students for "...lifelong learning, for active roles in a multiracial and multicultural world"; and outreach and public service activities undertaken by the University are to assist "...individuals, institutions, and communities in responding to a continuously changing world." (2) The global mission of the University is further supported by the designated themes of the Liberal Education Requirements, which include "International Perspectives." (8)

The University of Minnesota College of Liberal Arts requires undergraduates seeking B.A., B.F.A., and B.I.S. degrees to demonstrate proficiency in a foreign language as one of their graduation requirements. The CLA bulletin explains the process for satisfying this requirement before students enter the College. It also states the following about this requirement: "The study of a second language is considered essential for a liberal education. CLA expects students to begin second language study in high school or earlier." (7) For students who need to satisfy the requirement following enrollment in CLA, the Graduation Proficiency Test is available in over twenty languages. This requirement was implemented in 1986 based on the recommendations of the first CLA Task Force on Foreign Language Instruction (Report 1983). The second CLA Second Language Task Force (Report 1991) made further recommendations concerning the requirement, administration procedures for the GPT, and credit policies concerning language courses.

History of the GPT

The First Task Force was charged by Dean Lukermann with making recommendations to guarantee quality education and to encourage students to arrive at the University prepared for college workloads (Report, 216). The Task Force Report found that better articulation was needed between regional high schools and the University to improve the circumstances under which students learn foreign languages. It also recognized that a "seat-time" requirements based on credits or courses completed was less effective than a proficiency-based requirement, which reflected actual student achievement and ability to use the foreign language. This finding took into consideration the diverse educational backgrounds of high school students, University undergraduates, and transfer students.

Following the recommendations made by the First Task Force, development of the GPT began in consultation with University faculty, language course instructors, and high school teachers. Initially the tests were incorporated as part of language courses, but this system proved to be administratively complex. The Second Task Force, which was formed by Dean Craig Swan, recommended that the Language Center coordinate administration of the GPT and that the GPT not be incorporated into sixth quarter language courses (Report, 236). The report concluded that the logistics involved in administering the test at the time made it undesirable for both instructors and students to incorporate the GPT into University language courses. Factors cited in the report as significant considerations included the potential loss of instructional time, the impact of testing on non-CLA students enrolled in the courses, test security, and heightened stress levels for students when administration of the test was spread out over a week’s time.

The report also emphasized the beneficial effects the GPT had had on student and instructor performance (235) Prior to implementation of the proficiency-based language requirement, student apathy in courses was high and class sizes had swelled to unmanageable proportions. Today the GPT gives students a meaningful goal to reach at the end of two years of language study. Class size has also been reduced to create instructional conditions appropriate to this requirement. Vertical articulation with high schools, other educational institutions, and upper division course has been improved while horizontal articulation across language departments at the University has been facilitated by the existence of the graduation requirement. The University of Minnesota is regarded as a national leader in the area of language teaching and has pioneered proficiency testing with support from federally-funded grants. Other colleges and universities in Minnesota and the U.S. that have language requirements based on proficiency tests include St. Olaf College, the University of Pennsylvania, the Residential College (University of Michigan).

Profile of the GPT

The Graduation Proficiency Tests are designed to measure the student’s ability to use the foreign language in a communicative context. While some differences exist in the tests used by the various languages due to language-specific differences, in general the GPT includes written sections (reading, writing, and listening) and an interview. The test is rated by instructors who have received training in the use of the test and are familiar with the grading rubrics established for the GPT. Interviews are tape recorded and double rated to ensure rating reliability. The language departments regularly review the results of student performance on the GPT, the design of the tests, and language course curriculum. The original test was developed by a committee of educators and specialists in Second Language Acquisition. The GPT is currently being revised through CARLA (Center for Advanced Research on Language Acquisition); as specified by federally-funded portions of the budget for this project, will eventually reach a national market.

The test allows flexibility for outstanding students. Since the requirement is based on language proficiency, rather than seat-time, students who excel in language classes and wish to take the GPT early may do so, and thus save tuition money. As in work settings, achievement has tangible rewards. For the rare instances in which students take two years of language courses prior to attempting the GPT and do not pass, the test also provides flexibility. Instead of repeating an entire course to achieve a passing grade (as would be the case if languages followed a seat-time requirement), the student is able to retake portions of the test as needed at any testing date throughout the year. Efforts to develop a fully computerized version of the written portions of the GPT for the Most Commonly Taught Languages (French, German, and Spanish) will eventually make it possible for students to take the tests on demand, adding additional flexibility.

Cut-off scores for the GPT are set at the level of achievement expected of students working at the C- (70%) level or better in language courses at the end of two years in University courses. It is a proficiency test rather than an achievement test. This means that although it is a high-stakes test, it tests language skills in context at a set level; achievement tests are designed to measure the maximum abilities of students. Briefly explained, students need to achieve a passing score on the GPT, but whether or not they achieve a high or low score does not affect their ability to pass the test if they have met or exceeded the cut-off score. Most GPT’s include three written sections that test writing, listening, and reading skills, and an oral interview. Students can use a strong performance in one or more of the written sections to offset a weaker performance in another area or areas. All students must past the interview portion of the test. Final exams, by contrast, do not in most cases undergo the lengthy process of piloting and statistical verification that has been applied to the GPT; they focus on specific course content, test the student’s knowledge of selective materials, frequently do not include an oral interview, and often do include grammar sections. Students who excel in speaking skills or who have difficulty with grammar concepts may find final exam formats more challenging than the GPT. Others may show "hot house" language abilities and perform better than expected if they are tested on content that is especially familiar, although they may be unable to use the language well in other areas. Student achievement on a final exam does affect the grade in a language course. A small number of students, in fact, pass the GPT, yet fail the language course in which they are enrolled. Anecdotal reports on these cases from instructors indicate that this outcome reflects an effort on the part of students to pass the test, followed by a decision not to complete full course requirements satisfactorily.

The majority of students who take the GPT pass the examination on the first attempt and students who have a certified disability (including documented test anxiety) are eligible for appropriate accommodations. A recent dissertation concluded following detailed analysis of test results that student choice plays a key role in success rates for the GPT (Eden). When student scores were examined, the following factors showed a correlation with the success rates for the GPT: Students who delayed language study to their final years at the University rather than addressing this requirement as freshmen (as they are advised to do), students who interrupted the sequence by taking time off from language courses, students who attempted the test before they have completed the language sequence, and students who did not take all parts of the exam or delayed taking the exam although enrolled in courses show a higher rate of failure. Annual reports concerning student petitions for an academic waiver of the GPT (handled by CASL) provide further confirmation of these findings.

More detailed information about the passing rates for the GPT is provided in data analysis conducted by Dr. Monica Eden, the current Director of Testing. Among the group of students who failed the GPT during 1106, obvious patterns emerged in Eden's (1998) research. The most significant factors were articulation and attitude. The majority of GPT fails began their University language education in 1101 and took the sequence at the extreme end of their college career, thus repeating prior language study at a point farthest removed from their high school language courses. Overall, the longer students postponed their University language study, the lower they started in the sequence and the more likely they were to fail the GPT. Less than 2.5% of GPT fails initiated their college-level language study in their first year at the University. Additionally, around 40% of the total failing population skipped at least one academic quarter with no language course after starting the language sequence, and more than half of these students who skipped quarters had multiple breaks and for periods longer than seven months. Smaller student subgroups were also found with patterns of behavior that would have an adverse effect on language learning, such as failing language courses, skipping courses in the sequence, or registering for language courses under an S/N grade base. Out of all students who attempted the GPT, 0.7% were CLA students who failed the GPT and eventually left the University without a degree. Among all students who eventually passed the GPT, including both 1106 and non-1106 students, more than 97.6% of the students passed in one or two attempts, including incomplete attempts. These statistics are a significant testament both to the quality of language instruction at the University and to the ability of students to attain the required proficiency level. The vast majority of students demonstrated functional second language proficiency according to an expected timetable. Less than 2.4% required significantly more time to raise their language proficiency and also needed in excess of two attempts to pass the GPT.

Articulation Issues

The practice of articulation means that various kinds of institutions (high schools, colleges, and universities throughout the state) work together in educating students. Although different textbooks and curriculum are used in different places, there is a shared sense of what students need to learn in this effort. This common focus is made possible by the GPT, and by the existence of national standards, such as the ACTFL Proficiency Guidelines.

It is not yet known how the new State Graduation Requirements will impact foreign language learning at the K-12 level. Both the entrance and graduation requirement encourage students to develop foreign language skills prior to enrolling at the University of Minnesota. As the Second Task Force report noted, responsibility for introducing language training should be located in the K-12 system and an unbroken stream of instruction should exist from high school into more advanced courses in CLA. (232) The GPT and the CLA proficiency requirement pre-date the current National Standards movement and they are nationally recognized as progressive educational policies.

New policies approved by ComSLE concerning retrocredits are based on the expectation that students need to arrive at the University with introductory language preparation completed. It should be noted that since it is anticipated that increasingly students will present different educational backgrounds in foreign languages in terms of numbers of years and credits completed, interruptions in foreign language study, and instructional calendars (immersion, traditional, and block scheduling), the proficiency standard will become an even more significant measure of the actual achievement of students.

Value for Students

Successful completion of the GPT is a valuable form of certification for students. Because the test is based on the ACTFL standards, other educational institutions and prospective employers can easily evaluate how proficient students are upon passing this test. Language course instruction in first- and second-year courses is shaped by the proficiency. Course design emphasizes all language skills (speaking, listening, reading, and writing), as well as a knowledge of culture. This means that students are rewarded for developing a full range of language skills.

Instructional quality in all types of courses needs to receive continual attention. The GPT is a powerful means to accomplish this because both instructors and students understand the educational outcome that is sought. The professional development that Teaching Assistants receive in proficiency-oriented instruction and testing methods is part of the specialized education we offer graduate students, and this is extended to other language course instructors teaching at the University. Students can be assured that the class they take is conducted well because all instructors know that students must pass the GPT to graduate.

The existence of the GPT in over 20 languages has created a climate of instructional coherence within CLA. While different departments may have different curricula and in some cases language-specific approaches or methodologies, common consensus is possible about what current language instruction should involve. The language departments, in the spirit of their mission as part of a land-grant institution, strive to educate all students, not just a few, selected majors.

Resources:

Barnes, Betsy K., Carol A. Klee, and Ray M. Wakefield. "Reconsidering the FL Requirement: From Seat-Time to Proficiency in the Minnesota Experience." Challenges in the 1990s for College Foreign Language Programs, ed. Sally Sieloff Magnan. Boston: Heinle & Heinle, 1991. 55-69.

College of Literal Arts, University of Minnesota, Bulletin 1996-1999: "A World of Possibilities." Minneapolis: U Minnesota, 1996.

Eden, Monica Marie. "German at the University of Minnesota: A Case for Articulation and Accountability in a Proficiency-Based System," diss. University of Minnesota, 1998.

Minnesota Daily Archives (search: foreign languages),

http://www.mndaily.com/cgi-bin/swish-cgi.cgi

Articles:

13 Jan. 1997: Jose G. Gomez, "CLA languages are valuable"

18 Oct. 1999: Bryan Keogh, "Foreign language study programs gain popularity"

22 Jan. 1996: Betsy Barnes, "Letters to the Editor"

22 Jan. 1996: Chad Thomas, "Here's why you need a second language"

"Report of College of Liberal Arts Second Language Task Force" (June 1991), W. Phillips Shively, Chair.

"Report of the CLA Task Force on Foreign Language Instruction" (June 1983), Byron Marshall, Chair.

"Standards for Foreign Language Learning: Preparing for the 21st Century." May 11, 1998. http://www.actfl.org/htdocs/pubs/standards.htm

"What is Articulation?" August 6, 1999. http://carla.acad.umn.edu/art-defn.html

Prepared by Charlotte Melin in consultation with and on behalf of ComSLE

February 4, 2000